Quantitative phase field simulations of polycrystalline solidification using a vector order parameter

Tatu Pinomaa, Nana Ofori-Opoku, Anssi Laukkanen, Nikolas Provatas

Research output: Contribution to journalArticleScientificpeer-review

2 Citations (Scopus)

Abstract

A vector order parameter phase field model derived from a grand potential functional is presented as an alternative approach for modeling polycrystalline solidification of alloys. In this approach, the grand potential density is designed to contain a discrete set of finite wells, a feature that naturally allows for the growth and controlled interaction of multiple grains using a single vector field. We verify that dendritic solidification in binary alloys follows the well-established quantitative behavior in the thin interface limit. In addition, it is shown that grain boundary energy and solute back-diffusion are quantitatively consistent with earlier theoretical work, with grain boundary energy being controlled through a simple solid-solid interaction parameter. Moreover, when considering polycrystalline aggregates and their coarsening, we show that the kinetics follow the expected parabolic growth law. Finally, we demonstrate how this vector order parameter model can be used to describe nucleation in polycrystalline systems via thermal fluctuations of the vector order parameter, a process that cannot be treated consistently with multiphase or multi-order-parameter based phase field models. The presented vector order parameter model serves as a practical and efficient computational tool for simulating polycrystalline materials. We also discuss the extension of the order parameter to higher dimensions as a simple method for modeling multiple solid phases.

Original languageEnglish
Article number053310
JournalPhysical review E
Volume103
Issue number5
DOIs
Publication statusPublished - May 2021
MoE publication typeA1 Journal article-refereed

Fingerprint

Dive into the research topics of 'Quantitative phase field simulations of polycrystalline solidification using a vector order parameter'. Together they form a unique fingerprint.

Cite this